摘要
蓝点马鲛(Scomberomorusniphonius)是中国近海重要的大型中上层经济鱼类,精准预测蓝点马鲛渔场分布对渔业资源评估与管理有重要意义。本研究利用蓝点马鲛捕捞数据与卫星遥感海表温度(sea surface temperature,SST)和海表叶绿素浓度(chlorophyll-aconcertation,Chla)数据,构建了基于自动机器学习的蓝点马鲛CPUE预测模型,通过与XGBoost模型、随机森林模型和广义加性模型(generalized addictive models,GAM)对比,自动机器学习模型的确定系数(coefficientofdetermination,R^(2))分别提高了51%、107%和117%,均方根误差(rootmeansquarderror,RMSE)分别降低了15%、28%和32%。通过模型预测的蓝点马鲛CPUE,开发了渔场丰度指数,分析了渔场丰度时空变动规律。结果显示:蓝点马鲛渔场丰度高值区在春季由于受到SST的影响较大,呈现向北及向近岸移动的趋势,这种现象与蓝点马鲛索饵及产卵洄游路线一致;同时,蓝点马鲛渔场丰度高值区纬度重心的北移,也与气候变暖影响下蓝点马鲛索饵和产卵的适宜温度区变动有关。通过气候事件指数分析发现,Nino 3.4指数与蓝点马鲛渔场丰度高值区面积呈显著相关。结合RCP2.6、RCP4.5、RCP6.0、RCP8.54种情景,分别预测了2100年蓝点马鲛的CPUE分布变动,发现随着全球变暖,蓝点马鲛CPUE高值产区北移,并相比2010—2015年蓝点马鲛CPUE预测平均值,4种RCP情景下分别上升了0.1、2.2、2.41和17.3 kg/h。本研究结果可为中国近海经济鱼种的渔情预报研究提供参考。
The Scomberomorus niphonius is an economically important large-scale pelagic fish in the coastal waters of China.The accurate prediction of the fishing distribution of S.niphonius is of great significance for the assessment and management of fishery resources.The development of satellite remote sensing and artificial intelligence has provided strong technical support and convenience for fishery forecasting research.In this study,a S.niphonius CPUE prediction model based on automatic machine learning was constructed using S.niphonius fishing data,satellite remote sensing sea surface temperature(SST),and sea surface chlorophyll-a concentration(Chl-a)data.Through comparison with the XGBoost,random forest,and generalized additive models(GAM),the coefficient of determination(R^(2))of the automatic machine learning model increased by 51%,107%,and 117%,respectively,and the root mean square error(RMSE)was reduced by 15%,28%,and 32%,respectively.A fishery abundance index was developed through the CPUE predicted by the model,and the temporal and spatial variation laws of fishery abundance were analyzed.The results showed that the high-abundance area of S.niphonius presented a trend in moving northward and nearshore because of the influence of SST in spring,which is consistent with the feeding and spawning migration route of S.niphonius.The northward shift of the latitudinal center of gravity in the high-abundance area for the S.niphonius fishery was related to the change in the suitable temperature area for feeding and spawning of the S.niphonius under the influence of climate warming.Through the analysis of the climate event index,it was found that the Nino 3.4 index was significantly correlated with the area of high abundance of S.niphonius fishery(r=0.58,P<0.05).Combined with RCP2.6,RCP4.5,RCP6.0,and RCP8.5 for the four scenarios,the distribution changes in the CPUE of S.niphonius by 2100 were predicted.It was determined that with global warming,the high-value production area of CPUE of S.niphonius will continue to move northward,and increase by 0.1 kg/h,2.2 kg/h,2.4 kg/h,and 17.3 kg/h,respectively,compared with the predicted average value of CPUE of S.niphonius from 2010 to 2015.The results of this study provide a reference for predicting the fishing conditions of economically important fish species in the coastal waters of China.
作者
秦元哲
周振佳
刘阳
田永军
程家骅
刘旭东
张勇
QIN Yuanzhe;ZHOU Zhenjia;LIU Yang;TIAN Yongjun;CHENG Jiahua;LIU Xudong;ZHANG Yong(Fisheries Oceanography Laboratory,Fisheries College,Ocean University of China,Qingdao 266003,China;East China Sea Fishery Research Institute,Chinese Academy of Fishery Sciences,Shanghai 200090,China;Shanxi Dadi Xinya Technology Co.,Ltd.,Taiyuan 030021,China)
出处
《中国水产科学》
CAS
CSCD
北大核心
2022年第9期1375-1387,共13页
Journal of Fishery Sciences of China
基金
国家自然科学基金重点项目(41930534)
中央高校基本科研业务费专项(202242001)
关键词
蓝点马鲛
渔情预报
机器学习
气候变化
渔场丰度指数
Scomberomorus niphonius
fishery forecast
machine learning
climate change
fishery abundance index
作者简介
秦元哲(1997-),男,硕士研究生,研究方向为渔业遥感.E-mail:qinyzouc@163.com;通信作者:刘阳,教授,博士生导师,研究方向为渔业遥感.E-mail:Yangliu315@ouc.edu.cn
